The long term goal is to describe how homologous recombination is initiated, how recombination intermediate structures develop, and how they, in turn, are resolved, in eukaryotic cells in culture, using the simple, well-characterized, replicatable and recoverable genome from adenovirus. A complete description of the various stages requires that both the DNA structures and the polypeptide functions involved be determined and their interactions elucidated. This long term goal is important both for an understanding of the recombinational potential of a human pathogenic virus and also of the capacity of the cultured cell to permit homologous recombination, description of which may be necessary for replacement 'gene therapy'. The specific aims of this proposal are to use viral infection and DNA-mediated transfection to determine whether or not early viral functions are involved, either directly or indirectly, in any of the three forms of recombination described. In the DNA-mediated overlap transfection studies, the main aims are to determine whether or not recombination proceeds via the formation of heteroduplex DNA, perhaps initiated at the artificial termini created by restriction endonuclease cleavage and, if heteroduplex DNA is produced, how it is resolved. In marker rescue, the aims are to determine the extent of concerted transfer of genetic information from donor to recipient and to discriminate between a single-strand and double-strand transfer mechanism. Finally, an attempt to detect repair of double strand gaps, analogous to that described in yeast, is described, using gapped SV40 molecules and either the adenovirus-SV40 hybrid Ad2+ND1 on the resident SV40 sequences present in COS cells as source of the repaired information.